Re: xAP Asterisk integration

[Gregg Liming <gregg@xxxxxxxxxxx>]

Paul Gale wrote:

> Is anyone in a position to be able to look at writing some xAP support
> for Asterisk?
>
> I know Patrick Lidstone did something a while ago but he seems to have
> disappeared without trace L
>
Hi Paul,

Since no one else has responded, I'll give it a try.  Can you
elaborate what you want/expect in the way of "...support for
Asterisk"?
I'm thinking along 2 categories: (1) function and (2) interface.  CID is
a rather obvious example of function.  Call logging and VMI notification
might be another.  The second category is specific to asterisk: AGI or
manager interface.  IMO, the issue here is both what it possible and
what level of control the user wants (or, just as aptly, doesn't want).
I'm rather assuming that as an asterisk user that you're familiar w/
both and their pros/cons.

I currently use a number of AGI (perl) scripts to implement various
aspects of HA--notably CID, TTS (back through the * stream) and a
directed command (through IVR menus).  Most of these are variants that
various Misterhouse (mh) users have contributed.  The CID script is the
least mh-specific as it tries to be as compliant to the meteor schema as
possible (although, there is additional information that is needed and
therefore a departure exists).  The TTS is not xAP at all (it streams
from one of the 2 cepstral engines).  The directed command is a
mh-specific xAP schema.  In addition, there exists a manager interface
application that does call logging (rather primitively) via xAP.  The
problem, however, w/ manager interface apps is the absence of context
information; so--if you happen to leverage macros or cascaded contexts,
then this approach is somewhat lacking.  So far, I've not modified any
of these application/scripts beyond what I personally need from them.
I'll be happy to share, however, if there's an interest.  It's a fair
assessment that the work to date was on getting things to "work"
and not
full compliance w/ all pertinent specs.  Specifically, the
implementations were done before the current schemas had matured (and
are still maturing)--most notably, the CID/Telephony related one(s).

Gregg